Kinetics of oxidation of cobalt(II) complexes of ethylenediamine

Kinetics of oxidation of cobalt(II) complexes of ethylenediamine tetraacetate and cyclohexylenediaminetetraacetate by halogens. William H. Woodruff, B...
6 downloads 0 Views 616KB Size
Inorganic Chemistry, Vol. 13, No. 11, 1974 2573

Oxidation of Con-EDTA and -CyDTA Complexes

Contribution from the Department of Chemistry, Purdue University, West Lafayette, Indiana 47907

Kinetics of Oxidation of Cobalt(I1) Complexes of Ethylenediaminetetraacetate and Cyclohexylenediaminetetraacetate by Halogens WILLIAM H. WOODRUFF, BARBARA A. BURKE, and DALE W. MARGERUM*

Received November 29, 1973

AIC30866C

The rate of oxidation of Co"LZ' complexes, where L is EDTA4' or CyDTA4- (truns-l,2diaminocyclohexane-N,N,N ',N'tetraacetate), by iodine or bromine is first order in Co1'L2- and first order in total halogen (X, + Xi). The resolved rate sec-' at 25") for the EDTA complex are 5.2 X for I,, 3.1 X lo-' for I;, 7.3 X lo-' for Br2, and 5.4 X constants (M-' lo-' for Br;. Those for the CyDTA complex are 6.4 X lo-' for I,, 2.2 X lo-' for I;, 2.0 X lo-, for Br,, and 1.1 X l o - , for Br;. The reactions follow an inner-sphere electron-transfer mechanism which takes place in oneelectron steps. Steric effects in the aminocarboxylate ligand account for the generally smaller rate constants for CO'*(CYDTA)~-compared to Co''(EDTA)*- with given oxidants. The AH values are smaller for bromine than for the iodine reactions but are similar for X, and X;. The A S values are negative for both the I; and Brireactions with CO"(CYDTA)~-.

*

*

Introduction An earlier report from these laboratories' described the kinetics of the oxidation of Fe"(CyDTA)2- by iodine and triiodide ion. The rate of oxidation by 1, is fast (the second-order rate constant is 1.24 X lo6 M-' sec-' at 25.0'). 'The I; rate is 250 times slower than the I2 rate at 25' but the activation entropy for the I; oxidation is unusually positive. In the present study the reactions under investigation are given in eq 1, where L is EDTA4- or CyDTA4- and 2c0''L2-

+

[ i"1x, + x-

2Co"'L-

+ 3x-

X is iodine or bromine. These cobalt reactions are eight to ten orders of magnitude slower than the corresponding reactions of FenL2-. Individual rate constants are resolved for the X2 and X; reactions. With the cobalt(I1) complexes the X; ion is only 1.3-16.7 times slower to react than X2, in contrast to the iron(I1) system. The activation entropies for I; and Br; are negative for the CO"(C~DTA)~reaction whereas the A S for the I; reaction with Fe"(CyDTA)2- is large and positive. In a subsequent report2 additional cornparisons of metal centers and ligands provide a basis for development of a general mechanism for reactions of this type. The present paper concerns the mechanism of the Co" oxidations and the effects of bromine vs. iodine and of CyDTA vs. EDTA.

*

Experimental Section Reagents and Materials. The sodium salts of Co1'LZ' (L = EDTA, CyDTA) were prepared by a modification of the method of Schwarzenbach and par is saki^.^ In the present preparation, 0.2 mol each of reagent grade H,L and COCO, were stirred together in 400 ml of H,O. When CO, evolution ceased, the solution was neutralized with solid NaOH and filtered. The fitrate was evaporated on a steam table until crystals appeared on the surface of the solution and then cooled to room temperature. The product was precipitated by slowly adding 600 ml of absolute methanol, filtered, and then dried overnight in a vacuum oven at 100". Good microanalyses were obtained for Na,Co%yDTA.3H2O and Na2ConEDTA.2H,0. Reagent grade sodium perchlorate was recrystallized twice before use. Cesium triiodide, obtained from Alfa Inorganics (Beverly, Mass.), was the source of iodine for the Co"L'-oxidations. The purity of this reagent was confirmed titrimetrically. All other chemicals were reagent grade, used without further purification. (1) W.H.Woodruff, D. C. Weatherburn, and D. W. Margerum, Inorg. Chem., 10,2102 (1971). (2) W. H.Woodruff and D. W. Margerum, Inorg. Chem., 13, 2578 (1974). (3) G. Schwarzenbach and G. Parissakis, Helv. Chim. Acta, 41, 2425 (1958).

Bromine solutions were prepared from elemental bromine and just prior to use were standardized spectrophotometrically at the isosbestic point of Br, and Br; ( h 457 nm, E 98 A 4 - l cm-'). Instrumentation. Reactions were followed using either a Cary 16 spectrophotometer interfaced to a Varian G-2000 recorder or a Cary 14 spectrophotometer. Absorbance-time data were normally taken from the recorder chart, although for some experiments data were taken on-line with a Hewlett-Packard 21 15A computer as described el~ewhere.~Measurements of pH were made using an Instrumentation Laboratory 245 pH meter with a reference electrode containing saturated NaCl.

Results General Data. All of the reactions studied were first order in total halogen concentration and first order in cobalt(I1) chelate concentration. The rate constants due to the reactions of molecular halogen and of trihalide ion were resolved by evaluating the second-order rate constant at various total halide ion concentrations. This second-order rate constant is designated ko throughout this work and is defined such that -d([X2] T)/dt = ko[X,lT[ConL2-], where [X2lT = CX2l + [x;lf The hydrolysis of bromine to hypobromous acid is a potential interference in the study of bromine reactions. It has been found' that the rates of oxidation of cobalt(I1) complexes of EDTA and CyDTA are two to three orders of magnitude faster with HOBr than with Br2. In the present study, pH and bromide ion concentrations were chosen such that the HOBr reaction was negligible. The notation used to identify the rate constants due to X, and X; is k

Co1IL2- + X, --!+products Co"Lz-

+ X;

(2)

k

-4, products

(3)

with the reaction designated by k , having activation parameters AH,* and AS,*. Linear least-squares methods were employed both for the resolution of kl and k , (using the equations previously developed') and for determination of the activation parameters. Error limits of the resolved rate constants and the activation parameters are the standard deviations of these least-squares fits. ConL2--Bromine Reactions. First-order behavior in both Co''L2-and bromine was confirmed for these systems by carrying out separate reactions with first one and then the (4)B. G.Willis, J . A. Bittikoffer, H. L. Pardue, and D. W. Margerum, Anal. Chem., 42, 1340 (1970). ( 5 ) W. H. Woodruff, D. W. Margerum, M. J. Milano, H . L. Pardue, and R. E. Santini, Inorg. Chem., 1 2 , 1490 (1973).

2574 Inorganic Chemistry, Vol. 13,No. 11, 1974

Woodruff, Burke, and Margerum

Table I. Kinetics Data for the Oxidation of Co1'LZ- b y Bromineu

L = CyDTA 25 .O 25 .O 25.0 25.0 25.0 40.3 40.3 40.3 54.9 54.9 54.9

0.040 0.080 0.540 0.580 1.040 0.080 0.580 1.080 0.040 0.5 40 1.040

0.018 0.036 0.018 0.035 0.018 0.034 0.034 0.032 0.016 0.016 0.016

L = EDTA 25.0 25 .O 25 .O

0.120 0.520 1.020

U 1.1 = 1.1 (NaC10, [ C O ~ ~ L "=] 1.00 X

5.00

x 10-4~.

7.8 7.8 7.8

1.53 1.50 1.30 1.15 1.14 1.46 5.61 5.45 31.0 21.8 20.7

x 10.~

x

x

10-~ 10-3

59.4 56.8 55.1

KB~,[ Br-I

Figure 1. Plot of data used to resolve k , (intercept) and k , (slope) for the oxidation of CoII(CyDTA)'- by [ Br,] T at the temperatures studied.

+ NaBr), pH 3.5-4 (0.1 M acetate buffer),

r

L = EDTA

'0

'-

'-

'-

'-

'-

(6) G. Schwarzenbach, Helv. Ckim. Acta, 3 2 , 839 (1949). (7)A. W. Adamson, R e d . Truv. Chim. Pays-Bas, 7 5 , 809 (1956). (8) R. Dyke and W. C . E. Higginson,J. Ckem. SOC.,1998 (1960). (9) B. Grossman, Ph.D. Thesis, State University of New York at Buffalo, 1969. (10) D.B. Scaife and H. J. V. Tyrre1,J. Ckem. Soc., 386 (1958). (11) W. H.Woodruff and D. W. Margerum, Inorg. Chem., 12, 958 (1973).

I

I

X = Bromine

-

other reactant in excess. Thereafter, the kinetic runs were performed under first-order conditions with bromine in excess. The values of ko were calculated from the apparent first-order rate constants, kapp,by the relationship ko = kaP,/2 [Br,] T. Solutions containing bromine, sodium bromide, and sodium perchlorate were mixed with those containing CoIILz-, acetate buffer, sodium perchlorate, and if necessary sodium bromide. The reactions were monitored at 540 nm for L = CyDTA and at either 540 or 590 nm, , A,( for [Com(EDTA)Br] '7for L = EDTA. The rates are independent of acidity for the pH range (3.0-4.0) used. Earlier workers6" have observed that the initial product of the oxidation of C0"EDTA'- by bromine is [Co"(EDTA)Br] '-. This is also the case in the present work. The observed product of the oxidation of ConCyDTA2- by bromine is CoI'CyDTA-. This is in accord with the expectation that expels bromide ion much faster than [Co"(CyDTA)Br] [Corn(EDTA)Br] '-. The elimination of chloride from [Corn(CyDTA)Cl] is 5000 times faster than from [Conl(EDTA)Cl] 9 9 and the ring closure of [Co"'(CyDTA)OH] is 1O4 times faster than that of [Co"'(EDTA)OH] 2-.5 When the oxidation of C0"CyDTA'- is accelerated by using higher concentrations of bromine, [Corn(CyDTA)Br] is detected as a product. Thus, the kinetic data reported here refer to the same oxidative process whether L = EDTA or CyDTA. The observed kinetics data for the oxidation of Co11L2by bromine are presented in Table I and resolved as shown in Figures 1 and 2 . The values of k l and k2 at each temperature are given in Table 11. The data of Scaife and Tyrrel" were used for the temperature dependence of the stability constant of Br,. During the study of the oxidation of Co"EDTA'-, the dissociation reaction of the initial product [Co"'(EDTA)Br] '-.+ Co'IIEDTA' + Br(4 1 also was observed. Reaction 4 is accelerated by the presence of bromine." Conditions were adjusted so that reaction 4 did not interfere with the present oxidation measurements. Co"L'--Iodine Reactions. These reactions were observed

I

10-

M except at 54.9", where [Co''CyDTA'-] = -I

--'

I 1

m'

05-

I

m .

Y "

--+ x

,

I

5

I

I

I5

10

K,,3[Sr-I

Figure 2. Plot of data used to resolve k , (intercept) and k , (slope) for the oxidation of Co"(EDTA)'-by [Br,] T at 25.0'. Table 11. Resolved Rate Constants for the Reactions of Br, and Brs- with Co"L2- (JJ = 1.l)

25.0 40.3 54.9

16.85 14.9 13.5

L = CyDTA (2.0 f 0.4) X (9.6 f 0.2) X (3.67 f 0.07) X 10.'

25.0

16.85

(7.3

f

L = EDTA 0.8) X lo-'

(1.11 f 0.05) X (5.18 ?: 0.05) X (1.96 + 0.01) X 10"

(5.42

f

0.07) X

Table III. Rate Constants from Integrated Rate Expressions for the Oxidation of Co1'L2- by Iodinea

25.0 25.0 25.0 25.0

55.5 55.5

55.5 55.5 71.5 71.5 25.0 25.0 25.0 41.0 41.0

52.5 54.5 71.5 71.5 a 1.1 =

5.00 x 10-3 5.00~ 1.00 x io-* 1.00 x i o + 5.00 X 5.00 x i o + 1.00 X l o - * 1.00 x 1 0 - 2 1.00 X l o - ' 1.00 x 1 0 - 2

L = CyDTA 1.00 x 10-3 1 . 0 0 ~10-3 5.00 x 1 0 - ~ 5.00 x 1 0 - ~ 1.00 X 1.00 x 10-3 5.00 X 5.00 x 1 0 - 4 5.12 X 5.00 x 10-4

5.00 x 1 0 - ~ 1.00 X l o - ' 1.00 x i o - 2 LOO x io-, 1.00 x io-, 1.00 x io-' 1.00 X lo-' 1.00 X lo-' l.0OX l o - '

L = EDTA 1.00 x 10-3 5.20 X 1.00 x 10-3 5.08 x 1 0 - ~ 5.08 x 10-4 5.00 x io-, 5.00 X 5.00 X 5.00 X lo-,

0.050 0.100 0.050 0.100 0.050 0.100 0.050 0.100 0.050 0.100

2.3 x 1 0 - ~ 2 . 0 ~1 0 - ~ 2.9 x 1 0 - ~ 2.3 x i o + 6.6 X 6.4 x 5.6 X 5.5 x 1 0 ' ~ 3.6 X 3.4 x

0.050 0.050 0.100 0.050 0.100

5.4 5.1 3.6 4.3 2.3 2.1 3.2 1.7 1.3

0.100 0.050 0.050 0.100

1.0 (NaClO,), pH 5-6 (unbuffered).

x X

lo-'

x x x 10-~ x 10-~ X

X X

lo-'

lo-' lo-'

Inorganic Chemistry, Vol. 13, No. 11, 1974 2575

Oxidation of Con-EDTA and -CyDTA Complexes Table IV. Kinetics Data from Initial Rates for the Oxidation of Co1IL2-by Iodinea

ap

L = CyDTA

25 .O 24.9 37.1 37.1 37.1 54.3 54.3

1.00 x 2.00 x 2.00 x 2.00 x 1.00 x

lo-, 5.00 x 10-3 1.00 x lo-,

5.00 X lo-' 1.00 x 5 . 0 0 ~1 0 - ~ 1.00 x 10-3 5.00 x 10-4 1.00 x 10-3 5.00 X

25.0 25 .O 37.2 37.1 41.8 54.3 54.2

1.00 x 2.00 x 1.00 x 102.00 x lo-, 1.00 x 2.00 x 10-3 4.00 x

L = EDTA 4.8 1.00 x 10-3 9.9 1 . 0 0 ~10-3 9.9 5.00 x 1 0 - ~ 4.9 5.00 x 1 0 - ~ 4.9 5.00 x 10-4 4.8 2.50 X 2.4

lo-, lo-*

lo-,

4.8 x 9.9 x 4.9 x 9.9 x 4.8 x 9.8 X 4.8 X

5.00 X

x x x x x

4.3 x 6.9 x 4.4 x 7.6 X 4.5 x 8.1 x 4.7 x

10-4 10-4 1010-4 10-O

10-4

4.3 6.9 7.6 4.4

1010-4 10-

10-4 10-4

low4 10-4 10-4

x 10-4 X X

x

lo-"

4.4 x 10-

4.8 x 10-4 2.5 x 1 0 - ~

X

x 10-4

6.2 x 4.2 x 2.4 X 2.2 x 2.5 x 1.5 x 1.6 x

10-~ 10-~

4.1 x 3.5 x 1.0 x 1.8 x 2.5 x 8.7 x 7.3 x

10-4

1010-4 10-~ 10-3

10-4 10-3 10-3 10-~

== 1 .O (NaClO,), pH 5-6 (unbuffered).

Table V. Resolved Rate Constants for the Reactions of I, and I>-with Co"LZF

Temp, "C

K I ,a M - '

25.0 f 0.1 37.1 f 0.1 .55.1 f 0.5 71.5 f 0.5

768 550 354 235

25.0 f 0.1 37.1 f 0.1 41.3 t 0.4 53.9 f 0.7 71.5 f 0.5

768 550 488 358 235

Intercepts of Figures ?and 4, M-' sec-'

k,,bM-'sec'' L = CyDTA (6.4 f 1.1) x 10" (2.8 f 0.3) X lo-: (1.8 f 0.2) X 10L = EDTA (5.2 * 0.1) x 10-4 (1.8 f 0.4) X (3.1 f 0.3) x (8.7 f 0.9) x 10-3

(8.7

f

5.5) X l o - '

(1.7 f 0.2) x 1 0 - ~ 9.0 x 10-3 (5.9

f

1.8) X

(4.2 t 2.7) x (1.1 f 0.7) X lo-' 0.11

k2,CM-' sec-'

(2.2 f 0.1) x 107.6 * 10-5 d (5.63 0.08) X 3.2 x 1 0 - ~

*

(3.1 f 0.4) X 1.5 x 1 0 - ~d (1.7 f 0.9) X (2.0 i: 0.4) x 10-3 8.7 x

a From data of A. D. Autrey and R. E. Connick,J. Amer. Chem. SOC.,73, 1842 (1951); M. Davies and E. Gwynne, ibid.,74,2748 (1952); Chim. Ital., 90, 1068 (1960). b Calculated from initial-rate k , values and k , values in last column, this table. C Slopes and G. Daniele, GLVZZ. of Figures 3 and 4 unless otherwise noted. d Estimated from k , activation parameters. I

under first-order and under second-order conditions and by the method of initial rates. The initial rate measurements were possible because of the sluggishness of the iodine oxidations and were desirable in order to evaluate ko at low iodide ion concentrations ([I-] < [IZ]T). In the reactions which were run under first-order or second-order conditions, iodide ion was present in large excess over total iodine. Solutions containing &I3, NaI, and NaC104 were mixed with those containing Na2Co"L ,and NaC104. All the reactions were followed at 540 nm, a Amax for ComL-, and the pH of the reaction mixtures was between 5 and 6. Observed rate constants which were computed from the first-order and second-order data using the appropriate integrated rate equations are given in Table 111. In order to calculate - d( [I&)/dt from the absorbance in the initial rate measurements, the following values of the molar absorptivity (M-' cm-') at 540 nm were 305 for Co"'CyDTA-, 324. for ConIEDTA-, 5.4 for Co"CyDTA2- and for Co"EDTA"-. The stability constant of I; and therefore the absorbance due to vary with temperature. At each temperature, the A (molar absorptivity) necessary to calculate d( [Izl T)/dt was obtained from the intercepts of the absorbance curves at zero time. Table IV gives the initial rate data for the Co1'L2--iodine reactions. (12) E. Jacobsen and A. R. Sedmer-Olsen, Anal. Chim. Acta, 2 5 , 476(1961). (13) M. Mori, M. Shibata, E. Jq'QnO, and H. Nakajima, Bull. Chem. SOC.Jap., 29,887 (1956).

The total iodine and iodide ion concentrations in Table IV refer to the average values over the time of the initial rate measurement. Table V presents the resolved rate constants for the I2 and 1;oxidations. The k , values are the least-squares slopes of the plots shown in Figures 3 (CyDTA) and 4 (EDTA). However, the k l values were calculated by correcting the initialrate ko values (given in Table IV) for the k2 contributions at the existing iodide ion concentrations. More precise values of k l can be obtained in this manner than from the intercepts of Figures 3 and 4 because of the predominance of the k , path under the initial rate conditions. Nevertheless these intercepts are given with their standard deviations in Table V, to show that the experimental error of the intercept embraces kl in all cases. Discussion Table VI summarizes the rate constants at 25" and the activation parameters of the reactions under study. For a given Co1'L2- complex the k , and k, values are similar for X2 and X3-, despite the electrostatic repulsion expected for X3- and the less favorable free energy change in transferring an electron to X3-. A compensating effect is the stronger Lewis basicity of X3- compared to X,. Coordination of X3to cobalt(I1) will be more favorable and will tend to offset the more difficult electron-transfer reaction. As a result the AHl* and AH2*values are the same within experimental error. The similarity of the AS,* and AS,* values for the cobalt(I1) reactions is in marked contrast to the behavior of

2576 Inorganic Chemistiy, Vol. 13, No. 11, I 9 74

Woodruff, Burke, and Margerurn

Table VI. Rate Constant and Activation Parameter Data for the Oxidation of Co11L2-by X, ( k , ) and X3- ( k , p

Reactants CO%YDTA~-, I,, I ~ COIIEDTA*-,I,, I$CoIkyDTA", B r , Br,' CoI1EDTA2-, Br,, Br3a

Rate constants at 25.0" and p 0'1

I

k , , M W 1sec-'

A H , * , kcal mol-'

6.4 x 1 0 - ~ 5.2 X 2.0 x 10-3 7.3 x l o - , = 1.0-1.1 M

21.0 + 0.3 23 i 2 18.3 i 0.3

,

I

1

X = Iodine L = CyDTA

A S , * , eu -7.3 * 0.8 +1+6 -9.5 + 0.9

k , , M - ' sec-l

A H 2 * , kcal mol-'

A S , * , eu

2.2 x 3.1 x 1.1 x 10-3 5.4 x l o - ,

21.1 i 0.7 25 + 3 18.0 * 0.3

-9i.2 t5+9 -12i 1

to the relative A H " and A S " values between the one-electron reductions of bromine and iodine." Mechanism of the Oxidation of CoI1L2-by Halogens. The reactions under study are inner-sphere oxidations. Direct evidence for the inner-sphere mechanism is available for the oxidation of Co"L2-by bromine, in that the initial product of the oxidation is [Co'"(L)Br] No [COI"(L)T]~-product is observed in the oxidation of Co1'L2-by iodine, but tlie elimination of I- from the mixed product is expected 1.0 be much faster than the oxidation in this case. This exptectation results from the facts that the rate of elimination of Xfrom [Co"'(EDTA)XI2- is faster for Br- than for C1- ,' and the I2 oxidations are about two orders of magnitude slower than the Br2 oxidations. Therefore the [Col"(L)I]',- intermediate is not detected. However, the activation parameters suggest that the mechanism of the bromine and iodine oxidations are the same. The rate of these reactions is determined by a process subsequent to coordination of the oxidant to the metal center; ic., the reactions are not substitution controlled. This is known because the substitution reactions of Co1'L2- are many orders of magnitude faster16 than the redox reactions i,n the present study. The process which evidently controls tlne rate of the oxidation reactions is electron transfer (a term which, as used herein, includes the necessary reorganizations prior to electron transfer). A second candidate for the rate-determining process, atom transfer, can be ruled out o n the following basis. The activation energy, if an atom-transfer mechanism were followed, should be related to the homolytic bond energy of the halogen. If this were the case, the iodine oxidations should be faster than those by bromine. The opposite is true, which is consistent with a bona fide electron transfer mechanism. Although it is conceivable that the reac'rions could proceed via two-electron transfer, the facts are not consistent with such a mechanism. The observed yield of [CoIn(EDTA)BrI2from the bromine oxidation of ConEDTA2-is 100%. The maximum expected yield of this mixed complex, if a cobalt(IV) reaction intermediate is formed, is 50%. Therefore the reactions occur in one-electron steps. A mechanism consistent with the foregoing discussion is given in eq 5-8, where eq 6 is the rate-determining step (rds).

'-,

Figure 3. Plot of data used to resolve k , (intercept) and k , (slope) for the oxidation of Co11(CyDTA)2-by [I,]T at the temperatures studied. I

03

,

I

X = Iodine

L = EDTA

Figure 4. Plot of data used to resolve k , (intercept) and k , (slope) for the oxidation of Co"(EDTA)"by [I,] at the temperatures studied.

FenCyDTA- where A S l * is slightly negative while AS2* has a large positive value.' The effect of electrostatic repulsion would tend to cause AS2* to be more negative than ASl*by 10-20 eu14 and although the relatively high ionic strength ( ~ M1 ) will moderate the effect, other factors must compensate for the electrostatic repulsion to cause a less extensive reorganizational requirement in attaining the activated complex for X3-. Although Co'ICyDTA2- is a slightly better reducing agent than CoT'EDTA2-,the oxidations of Co"L2- by a particular oxidant are faster, by factors up to 50, when L i s EDTA than when L is CyDTA. The bromine oxidation rates are generally two orders of magnitude faster than the corresponding iodine reactions. Table VI shows that this effect is primarily due to lower A H * values in the bromine case. This is expected by analogy (14) A. A. Frost and R. G. Pearson, "Kinetics and Mechanism," 2nd ed, Wiley, New York, N. Y., 1961, p 145.

rapid

x . + x .-

x,

( 1 5 ) W. H. Woodruff and D. W. Margerum, Inorg. Chem., 12, 962 (1973). (16) J. P. Jones and D. W. Margerum, Inovg. Chem., 8, 1486 (1969).

Inorganic Chemistry,Vol. 13,No. 11, 1974 2577

Oxidation of Co'I-EDTA and -CyDTA Complexes The mechanism is written for the kl path but is equally applicable to the k2 path (the oxidation by X;). It is impossible to say, on the basis of the results of this study, whether the rapid process after the ratedetermining step is oxidation by the halogen radical (eq 7a) or radical recombination (eq 7b). However, halogen radical recombination reactions are known to be extremely fast .17$'* Effect of Steric Compression. The resolved rate constants for the oxidations of Co"L'- can be considered to consist of the product of the equilibrium constant for the formation of the intermediate (eq 5) and the rate constant, k,,, for the electron-transfer step (eq 6). Table VI1 compares the stability constants K and K for the reaction in eq 9 when L is ML t B

3 M(L)B kd

(9)

EDTA and CyDTA, respectively, for various identities of M and unidentate ligand B. The data presented are selected as representative of a larger body of available i n f ~ r m a t i o n 'and ~ serve to illustrate the following point. For metals of the approximate size of the first-row transition elements, the stability of M(L)B is one to two orders of magnitude greater for L = EDTA than for L = CyDTA, given a particular M and B. The ratio K/K' appears to be relatively constant despite changes in the charge of M or the magnitude of K. In CyDTA complexes, there is significant steric interaction between the methylene groups in the 3 and 6 positions of the cyclohexane ring and the methylene groups of two of the glycinate rings of the chelate.20'21 This effect is known as steric compression and is absent in EDTA complexes. Steric compression has been proposed to be responsible for the increased ring-closure rates of Co''€yDTA mixed complexes over those of the corresponding EDTA complexes?' The decreased stability of mixed complexes, evident in Table VII, when L = CyDTA is probably a manifestation of steric cornpression in the CyDTA chelates. The consequence of this ratio of stabilities is the prediction that the rate constant for the oxidation of an EDTA chelate will be one to two orders of magnitude greater than (17) L. I. Grossweiner and M. S. Matheson, J. Phys. Chem., 61, 1089 (1957). (18) A. T. Thornton and G. S. Laurence, Chem. Commun., 443 (1970). (19) L. G. Sillen and A. E. Martell, Chem. SOC.,Spec. Publ., No. 17 (1964);Suppl., No. 1 (1971). (20) J. L. Sudmeier and G. Occupati, Inorg. Chern., 7,2524 (1968). (21) J. L. Sudmeier, A. J. Senzel, and G. L. Blackmer, Inorg. Chem., 10,90 (1971). (22) M. H.Evans, Ph.D. Thesis, State University of New York at Buffalo, 1971.

Table VII. Comparison of Stability Constants, K , for MixedComplex Formation between ML and B (L = EDTA, CyDTA) M

B

K, M - ' (L =EDTA)

(L=CyDTA)

K',M.-' K/K'

c0(11) Al(III)c Cu(I1) Fe(II1) Ni(II)h

CNOHOH0,HCN-

2x 103~ 1.2 X 10' 251d 9.8 X l o 3 f 1.2 X l o 4

39b 2.5 X lo6 3.14e 742g 310

51 50 80 13 39

S. Nakamura, Ph.D. Thesis, University of Chicago, 1964. C G . Schwarzenbach, R. Gut, and G. Anderegg, Helv. Chirn. Acta, 37,937 (1954). d T. R. Bhat and M. Krishnamurthy,J. Inorg. Nucl. Chem., 25, 1147 (1963). e J. D. Carr, Ph.D. Thesis, Purdue University, 1966. f M. T. Beck and B. Csiszar,Magy. Kern. Foly., 66, 259 (1960). g M. Orhanovic and D. W. R. G. Wilkins,Croat. Chem. Acta, 39, 149 (1967). Margerum and L. I. Simandi, Proc. Int. Conf. Coord. Chem., 9, 37 a

b Reference 16.

(1966).

that for the oxidation of a CyDTA chelate if the oxidant occupies a normal coordination position in the inner sphere of the metal ion. As was noted earlier, this prediction is generally upheld for the oxidations of Co"L2- by halogens. Conclusions The evidence developed here indicates that the oxidations of Co"L2- by halogens proceed via an inner-sphere mechanism in which one-electron transfer is the rate-determining step. Steric compression effects are invoked to explain the relative rate behavior of the CyDTA and EDTA complexes. The X3- reactions are slower than the Xz reactions, but the better coordinating ability of X3- appears to offset its electrostatic repulsion and less favorable electron-transfer characteristics. Thus, the ratios of k l / k 2are not very large and equal 2.9,16.7,1.8,and 1.3 for the systems studied. The AH1* and AH,* values for a given system are the same within experimental error. However, the AH1 and AH2 values for bromine are 3 kcal mol-' smaller than those for iodine. The A&* values for the IC and for the Br3- reactions with Co"CyDTA'-are approximately - 10 eu in marked contrast to the large positive AS,* value ($28 eu) for the reaction of 1; and FenCyDTA'-. The oxidation of Co''CyDTA'- by I2 is 10 orders of magnitude slower than the oxidation of Fe'CyDTA2-by I*, a truly striking difference in rate for two so similar reactant pairs.

*

*

Acknowledgment. This investigation was sponsored by the Air Force Office of Scientific Research under AFOSR Grant 71-1988. Registry No. CoIIEDTA*', 14931-83-0; CoIICyDTA2-,2816191-3; I , , 7553-56-2; 1 3 - , 14900-04-0; BI,,7726-95-6; B I ~ - ,14522806.